From 660376d722ac1e258824d711977248896d255fcd Mon Sep 17 00:00:00 2001 From: HelloAnner Date: Sat, 11 Apr 2026 02:03:48 +0000 Subject: [PATCH] fix(testing_utils): check CUDA availability before get_device_capability Fixes #45341 When CUDA is installed but no GPU is available, testing_utils.py line 3223 calls torch.cuda.get_device_capability() which fails. Changed the condition to check torch.cuda.is_available() before accessing GPU capabilities. --- src/transformers/testing_utils.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/transformers/testing_utils.py b/src/transformers/testing_utils.py index 6e35a836db16..b0e5c941a77e 100644 --- a/src/transformers/testing_utils.py +++ b/src/transformers/testing_utils.py @@ -3198,9 +3198,9 @@ def get_device_properties() -> DeviceProperties: """ Get environment device properties. """ - if IS_CUDA_SYSTEM or IS_ROCM_SYSTEM: - import torch - + import torch + + if (IS_CUDA_SYSTEM or IS_ROCM_SYSTEM) and torch.cuda.is_available(): major, minor = torch.cuda.get_device_capability() if IS_ROCM_SYSTEM: return ("rocm", major, minor)